rr1076 - spontaneous ignition of gas turbine lubricants at ... · curves which illustrate how...
TRANSCRIPT
Spontaneous ignition of gas turbine lubricants at temperatures below their standard auto-ignition temperatures
Prepared by the Health and Safety Executive
RR1076 Research Report
© Crown copyright 2016
Prepared 2014 First published 2016
You may reuse this information (not including logos) free of charge in any format or medium, under the terms of the Open Government Licence. To view the licence visit www.nationalarchives.gov.uk/doc/open-government-licence/, write to the Information Policy Team, The National Archives, Kew, London TW9 4DU, or email [email protected].
Some images and illustrations may not be owned by the Crown so cannot be reproduced without permission of the copyright owner. Enquiries should be sent to [email protected].
This report and the work it describes were funded by the Health and Safety Executive (HSE). Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy.
There have been a number of incidents resulting in lubricating oil leaking in offshore gas turbine enclosures which could ignite if they came into contact with hot surfaces below their Auto Ignition Temperature (AIT). To assess the risk of auto-ignition, standard minimum AITs are used. However, AITs under industrial conditions are difficult to calculate and can be less than these standard values.
This report describes research using a Spontaneous Combustion Calorimeter developed to study spontaneous ignition. Preliminary tests were done for a range of process conditions that can influence minimum AITs for a number of gas turbine lubricating oils. These showed that ignition can occur at temperatures well below the standard minimum AIT. This indicates that if manufacturers rely on standard AITs at the design stage of gas turbines and enclosures, it may lead to a system that is likely to increase the ignition probability of any flammable release. To confirm these findings, further tests would be needed over a wider temperature range and under conditions which more closely represent the conditions in gas turbine enclosures.
Until AITs under industrial conditions are understood and addressed in design criteria, dutyholders will need to err on the side of caution in identifying and adequately controlling potential ignition sources.
HSE Books
Lorem ipsum dolor sit amet consectetuer adipiscing elit
Dr TJ Snee, R Braddock and Dr JT Allen Health and Safety Executive Harpur Hill Buxton Derbyshire SK17 9JN
Spontaneous ignition of gas turbine lubricants at temperatures below their standard auto-ignition temperatures
1
ACKNOWLEDGEMENTS
Acknowledgement is given by Dr T Snee, primary author, and Dr J T Allen of the extensive
practical and other assistance provided by Mr R Braddock throughout the course of this work.
This includes production of the design and build specification documentation contained within
this report as Appendix A.
2
EXECUTIVE SUMMARY
In order to determine how process conditions can influence minimum auto-ignition
temperatures in gas turbine enclosures, and other industrial installations, a novel type
of calorimeter has been developed at HSE's Buxton research laboratory. This instrument,
a Spontaneous Combustion Calorimeter (SCC), has been used to investigate the
ignition properties of pure substances, such as n-heptane and compare the results with minimum
ignition temperatures determined using standard methods. This was followed by a
more detailed investigation of the ignition of gas-turbine lubricants.
Standard minimum auto-ignition temperatures (A.I.T.s) are included in manufacturers’ safety
data sheets and are widely used to assess the risk of auto-ignition in industrial installations.
Minimum ignition temperatures under industrial conditions are less than the standard A.I.T.s
because increases in scale reduce the rate of heat loss per unit volume. It is difficult to calculate
the reduction in minimum ignition temperature associated with increasing scale. Many risk
assessment apply an arbitrary margin of safety between process temperature and the standard
A.I.T.
Results of experiments on n-heptane were consistent with previously published data and
confirmed that the SCC can be used to develop kinetic models for the slow oxidation processes
which lead to ignition.
The gas turbine lubricants were found to auto-ignite at relatively high temperatures > 350oC,
with no evidence of slow oxidation preceding ignition. The results indicated that aging had no
strong influence on the minimum ignition temperature of the lubricants.
It is generally assumed that the standard method of measuring A.I.T. yields values which are
lower than those obtained from alternative laboratory-scale procedures. The minimum ignition
temperature when fuel is added to a static, enclosed, volume of heated air in the standard
apparatus is lower than that obtained at a heated surface or when the fuel-air mixture flows at a
significant rate through the test vessel. However, results for the gas turbine lubricants in the
SCC indicate that, if cold air is added slowly to heated fuel (rather than fuel added to hot air in
the standard procedure), ignition can occur at temperatures well below the standard A.I.T. This
would imply that operators/Dutyholders may still have a significant fire risk even where they
have, in good faith, employed all suitable engineering and other controls to reduce the
theoretical risk to ALARP levels, based on published A.I.T.s.
The experiments on the gas turbine lubricants need to be repeated over a wider temperature
range and under conditions which more closely represent the conditions in the gas turbine
enclosures. If further experiments support the preliminary interpretation of the initial
experiments, the wider implications for other types of industrial installation should be
examined.
3
4
CONTENTS PAGE
1. INTRODUCTION .................................................................... 5
2. LABORATORY-SCALE APPARATUS .................................. 6
2.1 Standard Auto-ignition Temperature (A.I.T.) 6 2.2 Spontaneous Combustion Calorimeter (SCC) 6
3. N-HEPTANE ........................................................................... 8
3.1 Experimental procedure 8 3.2 Results 8
4. GAS TURBINE LUBRICANTS ............................................. 11
4.1 Experimental procedure. 11
4.2 Experimental results 11 4.3 Comparison with standard A.I.T. 14 4.4 Temperature dependence of Peak B 15
5. DISCUSSION........................................................................ 17
6. CONCLUSIONS AND IMPLEMENTATION .......................... 18
7. REFERENCES ..................................................................... 19
APPENDIX A - HSL SCC .............................................................. 20
A1 – Calorimeter Design and Build 20 A2 - Calorimeter control system 39
A3 - Crydon D2450 Solid state relay specification data sheet 43
5
1. INTRODUCTION
The work described in this report forms part of an investigation of combustion hazards in gas turbine
enclosures requested by HSE Energy Division (ED3.2) in response to a series of fires that have
occurred at offshore installations. Other work in this investigation (Fletcher J, 2014) has concerned
itself with identifying common causes or modes of failure, based on reports of previous fires, and
reviewing relevant advice available to turbine Dutyholders.
The objectives of the study covered in this report, which concern auto-ignition hazards, are:
a) the design and construction of a laboratory-scale apparatus, and
b) the development of an associated methodology for assessing the likelihood and consequences
of auto-ignition in gas turbines.
The reliability of the methodology was to be determined using substances with well-established
physical and chemical properties. A calorimeter, designed for the investigation of auto-ignition, has
been constructed at the HSE's Buxton research laboratory. After a review of the initial experimental
data on n-heptane, Energy Division (ED 3.2) requested further work on samples of used
turbine oil, obtained from a number of offshore installations, and a fresh sample of turbine
oil from the manufacturer.
Process temperatures which can result in auto-ignition are determined by the operating environment
and the physical and chemical properties of the process fluids. The heat transfer characteristics of the
process environment or hot surface can have a strong influence on the minimum ignition temperature.
In general, an increase in vessel volume will result in a reduction in auto-ignition temperature.
Ignition at hot surfaces requires a higher temperature than that necessary to produce ignition in a
closed vessel.
Minimum auto-ignition temperatures (A.I.T), measured in laboratory-scale equipment, are used both
by manufacturers to minimise the initial risk of ignition through the design process and also by
operators/Dutyholders to assess the residual risk of ignition in industrial processes. The standard
apparatus for measuring A.I.T. is considered in the first section of this report, followed by a
brief description of the new Spontaneous Combustion Calorimeter (SCC) developed at HSE's Buxton research laboratory. Detailed SCC design and build information is presented in Appendix A.
The performance of the SCC has been assessed using n-heptane. Results of these experiments are
reported and compared with data from an earlier version of the SCC. This discussion includes a set of
curves which illustrate how kinetic data can be obtained from the SCC and used to calculate minimum
process temperature that can lead to ignition.
The experimental procedure and results for the gas turbine lubricants are described in the main section
of this report. The results demonstrate how reliance by manufacturers on the standard A.I.T.s at the
design stage of gas turbines and their enclosures can lead to a system that is likely to increase the
ignition probability of any flammable release. Effectively, operators/Dutyholders may still have a
significant fire risk even where they have, in good faith, employed all suitable engineering and other
controls to reduce the theoretical risk to ALARP levels.
6
2. LABORATORY-SCALE APPARATUS
2.1 STANDARD AUTO-IGNITION TEMPERATURE (A.I.T.) In the standard methods for determining A.I.T., a small quantity of fuel is injected into a uniformly-
heated flask held in a thermostatic oven. Fuel is injected, over a range of oven temperatures, until
ignition is detected by the observation of a flame or a sharp increase in temperature inside the flask.
Measurements are repeated, with various amounts of fuel, in order to determine the minimum oven
temperature necessary for ignition. The standard methods use a conical or spherical flask ranging in
size from 125 to 250 ml.
The standard A.I.T. cannot be used directly to determine a safe equipment operating temperature
because of the effect of scale and the differences between process conditions and the conditions in the
standard apparatus.
Differences can arise due to, for example;
variations in the degree of agitation
non- uniform temperature and concentration distributions in the process vessel
effects of flow on the rate of oxidation of the fuel air mixture
variation in the geometry of the process vessel
reduced rates of oxidation in the presence of inert gas
In order to quantify these effects it is necessary to obtain experimental data under conditions which
can be related to the conditions in the industrial process. The standard measurement of A.I.T. needs to
be interpreted using measurements of the temperature and concentration dependence of the slow
oxidation reactions which precede ignition. The SCC has been designed to provide detailed data
which can be used to develop a kinetic model for the oxidation reactions. However, in many cases, it
is sufficient to use the instrument to investigate how variations in specific process parameters can
affect the conditions for ignition. In this way, it may be possible to determine the margin of safety
which needs to be established between the process temperature and the standard A.I.T.
2.2 SPONTANEOUS COMBUSTION CALORIMETER (SCC) The SSC is shown schematically in Figure 1. The sample is held in a thin-walled stainless-steel test
cell which is surrounded by guard heaters mounted around a copper tube. Adiabatic conditions are
achieved by matching the temperature of the guard heaters to that of the sample. The calorimeter can
also operate in isoperibolic mode, i.e. where the surroundings of the reaction mass are maintained at a
constant temperature so that any exothermic or endothermic changes produce a corresponding
temperature change in the reactor. Such operation is achieved here by setting the guard heat to
maintain a constant wall temperature. Peripheral equipment is provided for various modes of sample
injection and adjustment of the pressure and composition of gas and vapour in the test cell. The
control systems and detailed technical specification for the instrument is described in Appendix A.
7
Figure 1 - Schematic diagram of the Spontaneous Combustion Calorimeter.
Further detail is provided in Appendix A.
T
T
Air Vacuum
P
Copper Calorimeter Chamber
Hotplate Stirrer
Exhaust
Test Cell
T
T
T
T
8
3. N-HEPTANE
In order to evaluate the performance of the SCC, experiments were carried out on n-heptane. This
pure substance has a flash point of -4
oC and a standard A.I.T of 223
oC.
3.1 EXPERIMENTAL PROCEDURE The experiments were performed by first setting the instrument to maintain a constant wall
temperature (isoperibolic conditions) and allowing sufficient time for the sample-can to reach the set
point. The sample-can was then connected to a vacuum pump and a partial vacuum was maintained in
order to remove volatile contaminants. The can was then isolated from the pump and connected to a
small reservoir which was open to atmosphere and contained n-heptane. The sample was drawn into
the can, followed by sufficient air to restore atmospheric pressure and then the connection to the
reservoir was closed.
3.2 RESULTS Figure 2 and Figure 3 show the results of experiments at set points of 230 and 249
oC, respectively.
After sample injection, both experiments show a gradual increase in temperature with the pressure
remaining at the initial value of 1 Bara. This was followed by a sharp increase in pressure and
temperature, indicating that ignition had occurred. After ignition, the temperature returned gradually
to the set point, but the pressure remained above atmospheric. This excess pressure is attributable to
gaseous products of combustion.
Results from the two experiments are summarised in Table 1. The data must be interpreted carefully
because of the limitations in response time of the pressure and temperature transducers, relative to the
timescale over which processes are occurring, at the time of ignition. The more gradual changes in
temperature, during the induction period, can be monitored accurately and these changes provide an
indication of the rate of acceleration of the slow oxidation reactions that lead to ignition. Subject to
the outcome of this work, higher response rate transducers may wish to be considered.
Observation of ignition at the set point temperatures of 230 and 249oC is consistent with experimental
data obtained previously, under adiabatic conditions (Snee & Montserrat, 2010). Under isoperibolic
and adiabatic conditions, temperature increases are lower than would be observed in an isolated
system. This is due to the thermal dilution due to heat transfer to the sample-can. Thermal dilution
allows measurement of rates of slow combustion over a wider temperature range. The temperature
increase in an isolated system can be calculated using the adiabatic data along with calculated values
for the thermal capacity of the fuel-air mixture and the thermal capacity of the sample-can.
Table 1 - Experimental results for n-heptane.
Set point 230 249
Initial temperature (oC) 231.3 249.0
Maximum temperature (oC) 245.1 267.3
Temperature increase (oC) 13.8 19.5
Initial pressure (bara) 0.995 0.995
Maximum pressure (bara) 1.291 1.233
Pressure increase (bara) 0.296 0.238
Ignition delay (sec) 28.3 s 16.0 s
9
220
230
240
250
260
270
242000 243000 244000 2450000.0
0.5
1.0
1.5
200 micro litres heptane
isoperibolic: 230°C
Pres
sure
(bar
a)
Time (100 ms)
pressure
Tem
pera
ture
(°C)
sample temperature wall temperature
28.3 sec
sampleinjection
Figure 2 - Temperature and pressure changes during ignition of n-heptane in the SCC at 230oC.
240
250
260
270
53000 54000 55000 560000.0
0.5
1.0
1.5
200 micro litres heptane
isoperibolic: 249°C
Pres
sure
(bar
a)
Time (100 ms)
pressure T
empe
ratu
re (°
C)
sample temperature wall temperature
16 sec
sampleinjection
Figure 3 - Temperature and pressure changes during ignition of n-heptane in the SCC at 249oC.
Kinetic parameters for slow combustion reactions can be obtained from a series of experiments under
adiabatic conditions over a range of initial temperatures. These measurements have been performed
using an earlier version of the SSC as part of the previous project. Figure 4 shows results from a series
of experiments with heptane-air mixtures over the temperature range from 190 to 230oC. The results
0 10 20 30 40170
180
190
200
210
220
230
240
1.0
1.5
2.0
2.5
3.0
0 20 40
180
210
240
ignition
Tem
pera
ture
(°C
)
Initial temperature 190°C 200°C 210°C 220°C 230°C
Time (min.)
Fitted curves for autocatalysis
Pre
ssur
e (b
ara)
corresponding pressure records
10
in Figure 4 demonstrate how progressive increases in the initial temperature and the corresponding
rates of slow oxidation eventually lead to supercritical conditions and ignition. The temperature
profile during ignition is distinct from the s-shaped temperature-time curves observed under
subcritical conditions. Figure 4 shows that there was no substantial pressure increase during slow
oxidation. The rapid pressure increase (to 1.4 Bar) was only observed during ignition.
Figure 4 - Adiabatic slow combustion and ignition data for n-heptane. The fitted curves, plotted
in black, are derived from a kinetic model of the slow combustion processes leading to ignition.
Gas turbine fuels are likely to auto-ignite over a temperature range similar to the range over which
slow rates of oxidation of n-hexane can lead to ignition. It would be useful to obtain adiabatic data on
these fuels so that the potential hazards due to fuel leaks in the turbine enclosure could be assessed.
Kinetic analysis of the results for n-heptane in Figure 4 can be used to calculate the minimum process
temperature that can lead to ignition in a system. Alternatively, the results could also be used in
C.F.D. modelling of temperature and concentration distributions in industrial installations, although
these are beyond the scope of the current work. Regulatory authorities and industrial consortia may
utilise this type of analysis in order to raise awareness of the problems that can arise if a standard
A.I.T. is misinterpreted at the design criteria stage, and subsequently. Results for gas turbine
lubricants in the following section demonstrate how important findings can be obtained without full
kinetic analysis.
0 10 20 30 40170
180
190
200
210
220
230
240
1.0
1.5
2.0
2.5
3.0
0 20 40
180
210
240
ignition
Tem
pera
ture
(°C
)
Initial temperature 190°C 200°C 210°C 220°C 230°C
Time (min.)
Fitted curves for autocatalysis
Pre
ssur
e (b
ara)
corresponding pressure records
11
4. GAS TURBINE LUBRICANTS
Energy Division (ED 3.2) provided 3 samples of lubrication oil from working offshore platform
compressor turbines. The bottles were labelled C1, C2 and C3. The bottles contained used oil but it
has not been possible to establish the exact provenance of the samples or the duration over which the
oil samples would have been subject to elevated temperatures in the turbine enclosures.
Further, in order to allow comparative aging tests to be undertaken, HSE obtained a sample
of representative unused lubricating oil, AeroShell Turbine Oil 500, direct from the manufacturer.
The flash point of each sample was measured using a Seta Multiflash apparatus and following BS EN
ISO 3679-2004. The standard A.I.T. of one of the samples (C1) was measured following the
procedure set out in DIN51 794: Testing of Mineral Oil Hydrocarbons. Results of these
measurements are summarised in Table 2 along with additional data obtained from the M.S.D.S for
the new oil.
Table 2 - Summary of test results on oil Sample C1 when tested in accordance with BS EN ISO
3679-2004.
C1 C2 C3 AeroShell
Flash point (oC) 209.5 209.5 209.5 260
(typical)
AIT (oC
) 340 >320
Density (kg m3) 1.005
Vapour density (air =1) >1
(estimated)
Initial boiling point (oC) >280
(estimated)
Upper/Lower Flammability or
Explosion Limits (oC
)1 – 10 %
(V)
4.1 EXPERIMENTAL PROCEDURE The gas-turbine lubricants were tested using a procedure similar to that followed for the experiments
on n-heptane. Fuel concentrations and temperature set-points were chosen over ranges close to the
values at which ignition was observed in the standard A.I.T. apparatus. Initial experiments showed a
significant temperature increase when air was used to flush the sample can at the end of the test.
Further experiments were performed using controlled flow-rates of air or nitrogen to flush the can in
order to investigate the mechanism that may be responsible for these increases in temperature.
4.2 EXPERIMENTAL RESULTS
4.2.1 Used oil Figure 5 shows the results of an experiment in which 0.3 ml of lubricant (sample C3) was drawn into
the test can when the wall temperature had stabilised at 370oC. Immediately after sample injection, the
temperature increased rapidly to 390oC (Peak A) and then gradually returned to the initial value of
370oC. After a further 15 min., a slow flow of air was established through the test can with a pressure
of 1.25 Bara at the inlet. This resulted in a rapid increase in sample temperature to 410oC (Peak B).
The shape of this peak was similar to the peak associated with fuel injection.
12
The pressure trace in Figure 5 shows an initial increase from 0.63 Bara to 1 Bara as fuel and air are
drawn into the sample can after partial evacuation. The increase in pressure to 1.25 Bara, at the end of
the experiment, is due to the application of air pressure to give a controlled rate of flushing.
Figure 5 - Chart to show temperature and pressure measured by the SCC during tests on
sample C3.
Experiments were performed under the following conditions
Set point temperature: 350oC, 360
oC and 370
oC
Fuel injected: 0.1ml, 0.2ml, 0.3ml and 0.4ml
Wait time between fuel injection
and air injection: 5min, 10min. and 15min.
The experiments were designed to determine the significance of a range of factors which could affect
thermal stability. Results and tentative conclusions (in italics) are listed below:
1. Temperature
Two peaks (A and B) were observed for all experiments with a set point of 370oC
At 360oC, Peak B was observed for all experiments, but Peak A was absent, apart from small
temperature changes observed for experiments using 0.3 and 0.4ml of fuel
At 350oC only Peak B was observed.
The minimum ignition temperature when used oil was added to air in the SSC was 360oC
2. Wait time
In experiments in which only the wait time between fuel and air injection was varied, there
were no significant changes in the temperature increase associated with Peak B
The wait time had no significant effect on the thermal stability of the mixture in the sample can.
0
1
2
3
4
5
800 1000 1200 1400 1600 1800350
360
370
380
390
400
410
420
C3 0.3 ml 15 min waitTe
mpe
ratu
re (°
C)
Time (sec.)
temperature
Pre
ssur
e (b
ara)
pressure
fuelinjection
air injection
Peak A Peak B
wait time
13
3. Nitrogen injection (Shown in Figure 6)
Only a small temperature change was observed when nitrogen was injected, instead of air, at
the end of the wait time.
Peak B was associated with oxidative self-heating.
4. Source of sample
At the same set point, similar results were obtained for Samples C1, C2 and C3.
Aging of the lubricant had no strong influence on the thermal stability
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
4600 4700 4800 4900 5000 5100 5200 5300330
360
390
420
0 .2 ml Nitrogen Flush
Tem
pera
ture
(°C
)
Time (s)
temperature
nitrogeninjection
pressure
Pre
ssur
e (b
ara) fuel
injection
Figure 6 - Temperature and pressure records for an experiment using 0.2ml of Sample C1
where nitrogen, instead of air, was injected at the end of the test.
14
2
4
6
10 12 14 16 18 20 22350
360
370
380
390
400
410
420
Fresh LubricantTr
empe
ratu
re (
o C)
Time (min.)
temperature
Peak APeak B
fuelinjection
airinjection
Pre
ssur
e (b
ara)
pressure
Figure 7 - SCC data for the ignition of AeroShell 500 at an initial temperature of 360oC.
4.2.2 Fresh Sample Results for AeroShell 500 at an initial temperature of 360
oC are shown in Figure 7. The temperature
increase associated with Peak B in this experiment was higher than that observed with used samples at
360oC. This indicates that the minimum ignition temperature of the fresh lubricant in the SCC may be
less than the value of 360oC observed for the used samples. Further experiments are needed in order
to confirm this conclusion. The relationship between ignition in the SCC and values obtained from the
standard A.I.T. apparatus is discussed in the following section.
4.3 COMPARISON WITH STANDARD A.I.T. Fuel is injected into hot air in order to determine minimum ignition temperature in the standard
apparatus for measuring A.I.T. A similar procedure was followed using the SCC but improved mixing
is obtained by partial evacuation of the test can prior to sample injection. Partial evacuation also gives
better control of the ratio of fuel to air. Auto-ignition of Sample C1 in the standard apparatus occurred
at a minimum temperature of 340oC. With a smaller test cell in the SCC, the minimum temperature
necessary to produce ignition (associated with Peak A) was 360oC. The increased surface to volume
ratio of the SCC sample can, compared with that of the 200 ml conical flask in the standard apparatus,
is consistent with the observed increase in minimum ignition temperature. Published data on the
variation of A.I.T. with experimental conditions indicate a linear correlation between minimum
ignition temperature and the surface to volume ratio of the test vessel (Swarts D E, Orchin M, 1956).
Application of this correlation, using the surface to volume ratio of the SCC test can, indicates that the
minimum ignition temperature in the SCC will be 21.1oC higher than that observed in the standard
apparatus. This is consistent with the experimentally observed difference of 20 o
C between minimum
ignition temperatures measured in the SCC and the standard apparatus.
The flow of air used to flush the test can in the SCC was much less than that which would be used to
flush the conical flask in the standard apparatus. At the end of a test in the standard apparatus, rapid
flushing quickly reduces the temperature of the contents of the flask and no second peak is observed.
Experiments using a range of air-inlet pressures would provide an indication of the range of air flow-
rates which can lead to ignition. If the air flow through a gas turbine is interrupted at elevated
temperature, gradual resumption of ventilation could be hazardous.
15
4.4 TEMPERATURE DEPENDENCE OF PEAK B The results for the used lubricants and the fresh sample of AeroShell 500 indicate that ignition can be
induced by injecting air into heated fuel as well as by fuel injection into hot air. Further work is
needed to confirm this conclusion and establish whether this behaviour is restricted to gas turbine
lubricants.
62 64 66 68 70 72 74 76 78 80 82 84 86 88 90310
320
330
340
350
360
370
380
390
400
410
0.2 ml 10 min wait
Tem
pera
ture
(oC
)
Time (min.)
Initial temperature
350oC340oC330oC315oC360oC370oC
Peak A
Figure 8 - Temperature records from a series of experiments in the SCC when the initial
temperature was progressively reduced from 370oC to 315
oC. (Between each data set the time
axis has been shifted by 1 minute, in order to facilitate comparison of the results.)
Figure 8 shows the results for Sample C1 over a range of initial temperatures. The experiments with
initial temperatures from 315oC to 350
oC were performed with a sample volume of 0.2 ml. The
experiments at 360 o
C and 370oC had sample volumes of 0.4ml and 0.3ml, respectively. Peak A was
observed in these two experiments and this is associated with auto ignition when fuel is added to hot
air. Peak A diminishes significantly between 370oC and 360
oC and has not been observed at
temperatures below 360oC. By contrast, Peak B, which is associated with the addition of cold air to
the heated contents of the sample can, is observed at initial temperatures as low as 315oC. At 315
oC,
the temperature increase associated with Peak B is relatively large (29.7 o
C) and indicates that the
peak will still be observed at initial temperatures substantially below 315oC.
The chemical composition of the gas turbine lubricants has not been investigated. The data sheet for
AeroShell 500 indicates that the oils comprise a blend of synthetic esters and additives. Only
“estimated” and “typical values” are provided for the vapour density and flammability limits but the
data indicates minimum ignition temperatures in the SCC were obtained under very fuel rich
conditions. It is possible that slow addition of air may reduce the minimum ignition temperature but a
review of published data has produced no evidence to support this hypothesis. Under fuel rich
conditions, at elevated temperature, air slowly added to the hot fuel may react and generate heat at a
rate sufficient to exceed any heat losses (‘cooling’) caused by the introduction of the cold air, thus
resulting in a net increase in temperature (heat). This increase in temperature would increase the rate
of oxidation, which may ultimately lead to ignition. This would need to be studied further
theoretically and experimentally, but is beyond the scope of the current project.
16
Experiments using a range of air-inlet pressures, during flushing, would provide insights into the
physical and chemical processes associated with Peak B. Increasing air flow rates will ultimately
cause cooling and reduce the likelihood of ignition. Further experiments should be performed to
confirm that the rapid rates of temperature rise associated with Peaks A and B are due to flame
propagation.
Detailed evaluation of the experimental results in relation to the temperature distribution and pattern
of air flow in gas turbine enclosures is beyond the remit for the present investigation.
17
5. DISCUSSION
The mechanism associated with the oxidation of hydrocarbons varies according to their chemical
structure. Hydrocarbons with a carbon-chain length greater than 5 have relatively low A.I.T.s,
typically <300OC and with significant delay between sample injection and ignition. The results for n-
heptane have shown how slow combustion can be observed at temperatures well below the A.I.T.
For substances with A.I.T.s greater than 300oC, the delay between sample injection and ignition tends
to be relatively short, with no evidence of slow combustion preceding ignition. The results for gas
turbine lubricants display this type of behaviour. In the SSC the sharp increases in temperature
associated with Peaks A and B occur immediately after fuel and air injection, respectively. There is no
evidence of slow combustion, preceding the peaks, or at temperatures below the minimum ignition
temperature. Adiabatic operation, at these temperatures, would not provide useful kinetic data. In this
temperature regime, controlled flow of the fuel air mixtures through a heated tube can be used to
observe the rate of vapour phase oxidation. The SCC can be configured to provide this type of
measurement.
It must be emphasized that this discussion is based on a tentative interpretation of the origin of Peak
B. The results have been obtained from new equipment only recently commissioned. Repeat experiments must be performed in order verify or dismiss the current hypothesis.
18
6. CONCLUSIONS AND IMPLEMENTATION
6.1 A new type calorimeter for studying spontaneous ignition has been constructed at HSE's Buxton
research laboratory has been shown to produce reliable data on slow oxidation and ignition of n-heptane/air
mixtures.
6.2 The calorimeter has been used successfully to observe the effects of a range of process
conditions which can influence minimum ignition temperatures.
6.3 A preliminary investigation of ignition properties of gas turbine lubricants has
identified measurements and precautions which could reduce the likelihood of ignition in the turbine
enclosures.
6.4 The experimental evidence indicates that, in laboratory-scale equipment, ignition can occur at
temperatures well below the minimum ignition temperature in the standard apparatus for
measuring AIT.
6.5 Reliance by manufacturers on the standard A.I.T.s at the design stage of gas turbines and
their enclosures can lead to a system that is likely to increase the ignition probability of any
flammable release. Effectively, operators/Dutyholders may still have a significant fire risk even
where they have, in good faith, employed all suitable engineering and other controls to reduce the
theoretical risk to ALARP levels. This may go some way to explaining the continuing significant
levels of such events (Fletcher J, 2014), despite the introduction of recent guidance.
6.6 It is recommended that in order to minimise the likelihood of
ignition, manufacturers/operators/Dutyholders should ensure the adequate design, operation,
modification and maintenance of oil, lubrication and fuel systems’ integrity, which could
otherwise result in a flammable release and/or source of ignition.
6.7 Assessments of residual ignition risk, based on the standard A.I.T, should be checked
to determine whether the process conditions are similar to those which produced reduced
ignition temperatures on laboratory scale.
6.8 In pragmatic terms, until any ramifications of this work are confirmed and addressed
in revised design criteria, Dutyholders will need to better identify and adequately control
potential ignition sources. Specific areas may include exhaust manifolds and pipework, turbines
and their enclosures, and heating elements. In practice, however, a fire risk may arise wherever an
accidental release (liquid, vapour or aerosol) of oil, lubricants or fuel is able to impinge on a
sufficiently hot surface. Based on the work in this report, further studies may redefine what
temperatures qualify as ‘sufficiently hot’ but current standard A.I.T.s should not automatically be
used as a ‘safe’ guideline.
19
7. REFERENCES
Fletcher J (2014). An Examination of Incidents Involving Gas Turbines and the Guidance That
Applies to Them, HSL Report FP/09/11, 2014.
Snee T J and Montserrat Siscart J RJ (2008). Assessment of the Critical Conditions for Slow
Oxidation and Autoignition in Large Process Vessels.13th International Symposium on Loss
Prevention in the Process Industries, Brugge, 2010.
Swarts D E, Orchin M (1956). Vapour-Phase Oxidation and Spontaneous Ignition –
3579Correlation and Effect of Variables. U.S. National Advisory Committee for Aeronautics.
Technical Note 3579, 1956.
20
APPENDIX A - HSL SCC
A1 – CALORIMETER DESIGN AND BUILD The Offshore Gas Turbine auto-ignition project required a bespoke adiabatic calorimeter to be
built by HSE. This appendix details the design and build of that calorimeter – a Spontaneous
Combustion Calorimeter (SCC). It is the aim of this overall project to gain an understanding of the
potential for ignition of flammable oil mists formed by normal operation of offshore gas turbines
and aged turbine lubricant oils in hot environments. The SCC would allow both investigation of auto-
ignition processes and comparative testing of specific oil samples acquired from dutyholders who
operate offshore gas turbines, where these oils are used for turbine lubrication.
The calorimeter, using IKA Labworldsoft software, was designed to run in either isothermal mode or
adiabatic mode.
Figure A1.1 shows a high level block diagram of the key elements of the calorimeter.
Figure A1.1 - Block diagram to show the flow of information between key parts of the
calorimeter.
Figure A1.2 shows the calorimeter being prepared for a test run, with the top plate assembly, sample
reservoir and vacuum pump assembly removed from the main body of the calorimeter.
Figure A1.2 - Photograph to show the calorimeter in-situ, with top plate assembly removed.
Calorimeter
chamber/body
Solid state relay control
box & temperature
transmitters
Control & data
acquisition
software
Video recording
& monitoring
21
When fully assembled, the calorimeter body was secured in place between a top and a bottom
stainless plate using threaded bar and appropriate nuts. A pressure tight seal was achieved using
suitable gasket materials at the joints. The copper body of the calorimeter and heaters were clad in
glass fibre wool insulation to minimise heat loss from the heated chamber (Figure A1.3). Early trial
runs highlighted the importance of routing the electrical cables and thermocouple extensions to the
outside of the insulation in order to reduce insulation burning damage. In Figure A1.3 therefore, the
cabling was routed away from the metal work, and the small signal and mains voltage were separated.
Figure A1.3 - Photograph to show the calorimeter hardware fully assembled.
The main body of the calorimeter was earthed by clamping an earth tag to the bottom steel plate, to
which all metalwork was connected by the threaded bar. The earthing tag was connected to mains
earth via the solid state relay box.
A1.1 - Electrical connection After each test, all mains wiring was visually checked for signs of heat induced damage. Initial test
runs showed that vulnerable areas of wiring were the glass braided single core cables connecting the
heater bands, and their crimped ring connectors. Thermocouple operation was tested after each run,
in situ, using a calibrated handheld thermometer, and this was then verified by the data logger PC.
22
Figure A1.4 - Photograph of the unfused mains plug socket used to change between glass braided cable and PVC insulated mains power cable.
The calorimeter’s band heaters received switched mains power from the solid state relay / control box
via twin core insulated cable. The change over from normally insulated mains cable to 2 x single
cored glass insulated cable was achieved using an un-fused mains plug (Figure A1.4).
Glass insulated cable
PVC insulated cable from
Solid State Relay box
23
A1.2 - Heater Power Consumption For the 3 x band heaters and the 1 x top plate ring heater, the maximum electrical power consumption
(assuming all four heaters are powered for 100% of the PWM duty cycle) was calculated as follows
based on:
Band heater, where:
Max power output = 770 W
Mains voltage = 230 V AC
Ring heater, where:
Max power output = 550 W
Mains voltage = 230 V AC
Four heaters (three band heaters and one ring heater) were powered in parallel via solid state relay,
and thus the power consumption is additive. Maximum current consumption was calculated to be:
Therefore it was concluded that the four heaters could all be powered from one mains plug socket,
protected by a 13 amp fuse.
24
Figure A1.5 - Photograph of the solid state relay control boxes and their connections to IKA
Digital I/O controller.
Figure A1.6 - Photograph of a Crydom D2450 solid state relay. Each relay control box contains
4 D2450 relays.
The 230 V AC power to the 3 x band heaters and 1 x ring heater is provided via Crydom D2450 solid
state relays (Figure A1.6). A copy of the solid state relay specification is presented in Section A3.
One solid state relay was assigned to each heater, which could be configured to form a series of
independent closed loop control systems with temperature feedback. However for the planned
application, the calorimeter was configured to allow all three band heaters to respond to one
temperature feedback input. This is diagrammatically shown in Figure A1.7. The solid state relays
are housed within a ventilated plastic enclosure to isolate the relays from the operator and reduce the
risk of electric shock (Figure A1.5).
Solid state relay
control boxes Interface between the solid state relay control box and the
IKA I/O control.
25
The calorimeter carries out sample temperature tracking by the use of closed loop feedback control.
In adiabatic mode the control loop set point temperature is determined by the temperature of the
sample under test that is resident in the test can. The sample temperature is measured via a K-type
thermocouple which is designed into the test can. Feedback temperature is provided by a single K-
type thermocouple. It should be noted though that the calorimeter has the capability to run four closed
loop temperature control loops, one for each control loop.
The feedback control thermocouple was mounted on the side of the calorimeter body in the gap
between the band heater fasteners. For thermal conduction, the thermocouple junction was encased in
copper plate. A single thermocouple was located on the top plate copper disk which houses the ring
heater. Figure A1.7 shows the control loop method adopted. A larger annotated version of this block
diagram is available in Section A2.
A ‘global’ set point temperature can be manually presented to the calorimeter to instigate a thermal
runaway event or to establish a benchmark auto ignition temperature. The control and data
acquisition system can then be switched to sample temperature tracking as the reaction / auto ignition
event begins.
The Proportional, Integral and Derivative (PID) controller, as used by the heater’s closed loop control
systems, were set up as P-D controllers (proportional-derivative). A limited amount of controller
tuning was conducted, but there is scope to further refine these controllers to achieve critical damping
of the overall system. Currently the PID controllers have the P, I and D settings shown in Table A1.
Figure A1.7 - Block diagram to show control and electrical power flow through the
calorimeter.
26
Table A1 - The P, I and D values used by each control loop.
Controller ID Proportional (P) Integral (I) Derivative (D)
PID 1 1 1000000 10
PID 2 not used not used not used
PID 3 3 1000000 10
PID 4 not used not used not used
The “I” term was eliminated by making it a very large value. In practice this meant that the controller
fell short of the set-point. This was not an issue because the compound heating effects of the
neighbouring band heaters brought the system into the set-point temperature region with less over
shoot than if an “I” term was also used.
The “D” term of 10 was chosen for both of the controllers used, because tests showed that this value
provided an acceptable recovery time after a loop disturbance, which allowed the “P” values to be
increased further without suffering control loop oscillation.
The output from the PID controller is fed into a Pulse Width Modulation (PWM) function. This
software driven function calculates an “on” period in the operating time window, based upon the PID
controller output. The output of the PWM function is a percentage of the operating time window,
referred to as the duty cycle, where mains power is provided to the heaters. Using this function, as the
set-point temperature is approached, the power provided to the heaters is progressively reduced on
every cycle until the set-point temperature is met. The output from the PWM is fed to the solid state
relay via an IKA IO controller, which is a digital control panel providing a 0 V or 5 V signal. The
output of the solid state relay control box is a 230 V AC, under load.
A1.2.1 - Heaters & Hotplate Figure A1.8 show the model of band heater used and the specification, as stated by the supplier,
Omega Engineering UK. Each heater has an independent 230 V AC mains supply, provided via a
Crydom D2450 solid state relay.
Figure A1.8 - Photograph to show the type of band heater clamped around the calorimeter
body.
Specification of the heater shown in Figure A1.8:
Omega Engineering Part Number: DB-050772
240 V One-piece band heater, barrel diameter 5”, 770 W
Heater 1
Heater 2
27
The calorimeter has the capacity to use four band heaters. Currently only three band heaters are used.
Each band heater is made up of two heater elements, joined at one end by a semi-malleable metal. At
the other end a bolt fastener allows the band heater to be tightly secured around the calorimeter body.
Each side of the heater can be powered separately, but in this application, the two heaters have been
wired in parallel to allow them to be powered from a single solid state relay. Figure A1.9 shows the
wiring of a band heater.
Figure A1.9 - Photograph to show the band heaters wired to the solid state relay.
Figure A1.10 - Photograph to show the type of disk heater used in the top plate assembly.
Specification of the heater shown in Figure A1.10:
Omega Engineering Part Number: A-205/240
A Series Ring Heater 240v, 500w Single-Heat Element Chrome Steel Sheath
The calorimeter uses one Ring Heater, which is part of the top plate assembly. This provides the
required heating for the top of the calorimeter chamber.
Heater 2
terminals
Cables to solid
state relay
Heater 1
terminals
28
Figure A1.11 - Photograph to show the disk heater in-situ in the top plate assembly.
A1.2.2 - Stirrer hotplate The stirrer hotplate purchased from Sigma Aldrich has the description: “IKA RCT basic IKAMAG
safety control. Universal hot plate magnetic stirrer, 20 L, 1500 rpm, 230 v, 1c/s”.
The hotplate stirrer is controlled by the Labworldsoft control and logger software. The controls as
shown in Figure A1.12 below are inactive, and any change made using them is overridden by the
Labworldsoft program.
Figure A1.12 - Photograph to show the front control panel of the hotplate stirrer and its position
in the calorimeter assembly.
Ring Heater, in top
plate assembly
Cables to solid
state relay
29
Figure A1.13 shows the connection between the hotplate stirrer and the Labworldsoft program. The
figure also highlights the PT100 resistance thermometer used in conjunction with the hotplate stirrer,
which indicates the temperature in the lower area of the calorimeter chamber. It should be noted that
by design the hotplate stirrer is limited to 300°C.
Figure A1.13 - Photograph to show the PT100 resistance thermometer and Labworldsoft
control cable connection to the hotplate stirrer.
The hotplate temperature is monitored using a PT100 (resistance thermometer), as shown in Figure
A1.14. The PT100 probe is housed in a push fit hole which goes through to the centre of the
calorimeter bottom plate; it is in direct contact with the IKA hotplate stirrer plate surface.
Figure A1.14 - Photograph to show the position of the PT100 resistance thermometer in-situ in
the calorimeter hotplate.
PT100 temperature
measurement device
Communication to
Labworldsoft
PT100
measurement
position
30
A1.3 - Temperature Measurement Figure A1.15 shows the temperature transmitters located adjacent to the calorimeter. Temperature is
monitored at five locations inside the instrument. The sixth transmitter, as shown in Figure A1.15,
will cater for the closed loop control system for a bottom heater, if required. These are necessary to
monitor the temperatures in different areas of the calorimeter body so that the control software can
extend or reduce the “on” period of the heaters when tracking test sample temperatures. The
temperature measurements are part of the closed loop control system (see Section A2 - Calorimeter
Control System).
Figure A1.15 - Photograph to show the temperature transmitters used in conjunction with the
Type-K control thermocouples.
All of the temperature transmitters used by the calorimeter are ranged to operate optimally in the 0 –
500oC temperature region, and they all only accept K-type thermocouples.
The measured temperatures are transmitted back to the Labworldsoft control and logger software via
the IKA DC2 I/O modules (Figure A1.16). The 4-20 mA currents, as measured by the temperature
transmitters, are appropriately scaled by Labworldsoft before being recorded as actual temperature
measurements. The temperature transmitters all have in-built cold junction compensation.
31
Figure A1.16 - Photograph to show connection from the Labworldsoft PWM function, via the
Datacontrol I/O2 to the switch signal input terminals of the solid state relays, as show in Figure
A1.6.
At the clamp point of each heater, a thermocouple plate was secured tightly against the calorimeter
body. This allowed the control temperature in the respective calorimeter heating band zone to be
measured. In addition to the three band heaters around the calorimeter body, thermocouples were also
placed to measure the temperatures of the top plate, bottom plate and test sample (Figure A1.17).
Figure A1.17 - Photograph to show the positioning on the calorimeter body of the Type-K
thermocouples that are used in the feedback control system.
The thermocouples are held in positioned in a push tight hole in a separate copper plate that sits in the
fastening gap. This ensures good thermal conduction between the calorimeter body and the
thermocouple, therefore increasing the accuracy of the overall temperature measurement. The copper
plate was secured into place, using four bolts, through a threaded steel plate which applies force to the
copper plate, bending it to the contour of the calorimeter body (Figure A1.18). It is important that the
32
copper plate sits tightly against the calorimeter body and air gaps are not present, which can lead to
inaccurate temperature measurement.
Figure A1.18 - Diagram to show cross-section from above of the securing mechanism of the
thermocouples used to measure calorimeter body temperature. The hole represents the location
of the thermocouple.
A1.4 - Pressure Transducer Changes in test can pressure can be measured when both the inlet and outlet valves are closed. This
calorimeter measures pressure using a 0 – 5 Bar (Abs) transducer (Figure A1.19). The transducer is a
4-20 mA device and connects to Labworldsoft via a fuse protected 250 resistor circuit contained in
the instrumentation cabinet. The IKA DC2 data acquisition module measures the voltage between the
resistor and ground. The voltage is then ranged between 0-5 V before being logged.
Figure A1.19 - Photograph to show the position of the pressure transducer on the calorimeter.
The pressure transducer is positioned on the inlet side of the test can via a Swagelok tee piece.
Copper plate
Thermocouple hole Calorimeter Body
Steel plate
33
A1.5 - Calorimeter test chamber assembly The body of the calorimeter is a single piece of copper pipe which has been machined to width,
reducing the calorimeter’s heat capacity, thereby improving the heating response time. Figure A1.20
shows the dimensions (mm) of the copper pipe that was used for the test chamber.
Figure A1.20 - Diagram to show the overall dimensions of the copper pipe which was used to
form the calorimeter body.
Figure A1.21 shows the actual calorimeter copper pipe with the top plate assembly removed during
test preparation.
Figure A1.21 - Photograph to show the calorimeter inner cavity from above. This view is from
where the top plate, top heater and sample reservoir would normally be sited.
200
120
140
mm
160
34
A1.6 - Top Plate assembly The top plate assembly consists of:
1. A ring heater
2. A copper disk, with diameter 115 mm
3. Top plate thermocouple
4. Phi-TEC calorimetric test can
5. Swagelok connections on the inlet and outlets
6. Swagelok tee piece on the inlet line for connection of pressure transducer.
Figure A1.22 - Photograph to show the top plate, top heater and sample test can assembly.
The test sample is injected through a valve and tube feed through the insulated cavity and finally into
the test can. The test can contains an integrated type k thermocouple (Figure A1.22). The cans used
are the same as those used by the HEL Phi-TEC calorimeter1. The assembly, shown in Figure A1.22,
is lowered into the calorimeter test cavity shown in Figure A1.21.
Figure A1.23 shows the sample reservoir in which the oil under test is held immediately prior to being
injected into the test can. The sample reservoir is seated on a Swagelok ferrule which allows the
vacuum pump or compressed air line to be connected allowing the test can to be cleaned.
1 http://www.helgroup.com/reactor-systems/thermal-hazards-and-calorimetry/phitec-ii/. Viewed on 30/05/2013.
1.
2.
3.
4.
5.
6.
35
Figure A1.23 - Photograph to show the sample reservoir attached to the inlet pipe of the test
can.
Figure A1.24 shows the piped connections to the ‘hastelloy’ sample can for sample injection,
evacuation and flushing using compressed air. This is the same test can as used by the HEL Phi-TEC
calorimeter. In order to position the test can in the top plate assembly, the thermocouple mini
connector has to be removed so that the sheath can be passed through the top plate hole. Once in
place, the appropriate Swagelok fitting and ferrules are slid onto the thermocouple, before the mini
connector is re-attached. This process has to be repeated every time a test can is replaced.
Figure A1.24 - Photograph showing an example of the test can used in the calorimeter.
The inlet, outlet and thermocouple are pre-attached to the test can. Internally a stirrer bar is present
that be agitated by the hotplate stirrer.
36
A1.7 - Control & data logging software The calorimeter is controlled using a Labworldsoft program which interfaces using two IKA DC2
data acquisition modules and an IKA I/O controller. Figure A1.25 shows the flow of data between the
calorimeter hardware and the Labworldsoft program.
SSR Box 11Top plate heater
Band Heater 1
Band Heater 2
Band Heater 3
IKA Hotplate & Stirrer
PT100
SSR Box 12
SSR Box 13
SSR Box 14
LabworldsoftControl
Program
Drawn: R Braddock HSL 10/12/2012Version: 1.01
lilac
blue
white
black
l.greenbrown
pink
grey
Control I/O2Chan 4
Control I/O2Chan 3
Control I/O2Chan 2
Control I/O2Chan 1
IKA Digital I/O controllerP
Transmitter 1
Transmitter 2
Temperature TransmittersType-K (0-500 degC)
Transmitter 3
Transmitter 4
Transmitter 5
lilacblue
whiteblack
green
pinkgrey
orangered
yellow
l.green
brown
Test Cancf35
37
cf33
35
cf41
43
cf40
42
cf39
41
cf38
40
DC2-2Chan 3
DC2-1Chan 4
DC2-1Chan 3
DC2-1Chan 2
DC2-1Chan 1
DC2-2Chan 4
Figure A1.25 - Block diagram to show the interfacing between the hardware and software of the
calorimeter control system.
It should be noted that Figure A1.25 shows the system configured to use all four closed loop
controllers. In reality Band Heaters 1-3 and the top plate heater are all controlled by the PWM signal
provided by Control I/O2, channel 3.
Figure A1.26 shows the Labworldsoft program that controls the calorimeter hardware. The main
features of the program are:
Heat ramping - This feature is used to bring the calorimeter hardware up to operating temperature in
a controlled manner, minimising set-point temperature overshoot. This feature can be tailored for use
with adiabatic measurement so that the calorimeter is brought up to a temperature which is
predetermined to be under the auto ignition temperature of the substance under test.
Temperature tracking (adiabatic measurement) - In this mode of operation the measured
temperature of the substance under test is “followed” by the calorimeter’s heaters.
Calorimeter Hardware Control interface & software
PWM Control Signal
Temperature feedback
37
Temperature set-point (isothermal measurement) - In this mode of operation, a predetermined set
point is entered into the program. Using PWM, the heaters warm the calorimeter to this set-point
temperature. The accuracy to which the set point temperature is met is dictated primarily by the tuning
of the PID control loop filters. The PID loops are tuned as stated in Table A1.
The software uses the following functions in order to achieve the measurement techniques:
PID controller - The larger the proportionality factor P, the faster the system deviation settles.
However P should not be too large to prevent control loop oscillation. Any remaining control loop
oscillation can be reduced further by the integral portion I. The controller dynamic is significantly
improved by the differential proportion D. The proportionality factor P may be greater than for a pure
P or PI controller without the regulator circuit becoming unstable.
PWM function - The Pulse Width Modulation (PWM) function calculates an “on” period in the
operating time window, based upon the PID controller output. The output of the PWM function is a
percentage of the operating time window, referred to as the duty cycle in which power is provided to
the designated digital output channels.
Figure A1.26 - Screenshot to show the Labworldsoft program function module interconnection.
All temperatures and duty times are displayed on the screen for the user to see and respond to (Figure
A1.27). The data are also written to a comma separated text file which is time and date stamped and
is readily processed using Microsoft Excel or similar spread sheet or data processing software.
38
Figure A1.27 - Screenshot to show the Labworldsoft user interface.
On the Labworldsoft display, each heater control loop has an associated temperature, control voltage
and duty cycle indicator (Figure A1.27). There are two primary controls on the display, the “T set”
slider which allows the user to specify a set point temperature, and the “RPM set” slider which allows
the user to specify a required hotplate stirrer revolution requirement. All measured temperatures are
viewable in real time via the temperature versus time graph, labelled “y/t graph” in Figure A1.27.
39
A2 - CALORIMETER CONTROL SYSTEM
40
Heater
Equipment under
Test (EUT)
PID
Control
PWM (v)
0-100%
0-240v
(AC)
EUT follower (DegC)
Heater
Chamber wall Temperature 1 (DegC)
Mains voltage (240v AC)
Low voltage (0-24v DC)
Thermocouple (Type-K)
Set-point
(Spt1)
Isothermal Testing
Adiabatic Testing
PWM (v)
0-100%
PID
Control Disk Heater
Top disk Temperature 2 (DegC)
Figure A2.1 - Block diagram to show the control system principle to be used on the auto ignition calorimeter. V1.00
41
Heater
Equipment under
Test (EUT)
Set-point
(Spt1)
PID
Control
PWM (v)
0-100%
0-240v
(AC)
EUT follower (DegC)
Heater
Internal Temperature 1 (DegC)
PID
Control
PWM (v)
0-100%
0-240v
(AC)
Internal Temperature 2 (DegC)
… … … … Mains voltage (240v AC)
Low voltage (0-24v DC)
Thermocouple (Type-K)
Chamber Temperature 1 (DegC)
Alternative SP feeds
Figure A2.2 - Block diagram to show the control system principle to be used on the auto ignition calorimeter - Original independent control.
V1.01
42
Heater
Heater
Heater
Heater
Band heater max output 770W
Band heater max output 770 W
Band heater max output 770 W
Band heater max output 500W
Calculation of maximum power consumption
Power = Voltage x Current
W = V x A
43
A3 - CRYDON D2450 SOLID STATE RELAY SPECIFICATION DATA SHEET
Series 1 240 VAC • Crydom's signature family of solid-state relays. Ratings from 10A to 125A @ 24-280 VAC
• SCR output for heavy industrial loads
• AC or DC control
• Zero-crossing (resistive loads) or random-fire (inductive loads) output
PRODUCT SELECTION Control Voltage 10A 25A 50A 75A 90A 110A 125A 3-32 VDC D2410 D2425 D2450 D2475 D2490 D21110 D24125
90-280 VAC A2410 A2425 A2450 A2475 A2490 A24110 A24125
18-36 VAC A2410E A2425E A2450E A2475E A2490E A24110E A24125E
OUTPUT SPECIFICATIONS (1) Description 10A 25A 50A 75A 90A 110A 125A Operating Voltage (47-63Hz) [Vrms] 24-280 24-280 24-280 24-280 24-280 24-280 24-280
Transient Overvoltage [Vpk] 600 600 600 600 600 600 600
Maximum Off-State Leakage Current @ Rated Voltage [mArms] 10 10 10 10 10 10 10
Minimum Off-State dv/dt @ Maximum Rated Voltage [V/μsec] (2) 500 500 500 500 500 500 500
Maximum Load Current (3) [Arms] 10 25 50 75 90 110 125
Minimum Load Current [mArms] 40 40 40 40 40 150 150
Maximum Surge Current (16.6ms) [Apk] 120 250 625 1000 1200 1500 1750
Maximum On-State Voltage Drop @ Rated Current [Vpk] 1.6 1.6 1.6 1.6 1.6 1.7 1.7
Thermal Resistance Junction to Case (Rjc) [°C/W] 1.48 1.02 0.63 0.31 0.28 0.25 0.22
Maximum I2 t for Fusing (8.3 msec) [A2 sec] 60 260 1620 4150 6000 9340 12700
Minimum Power Factor (with Maximum Load) 0.5 0.5 0.5 0.5 0.5 0.5 0.5
INPUT SPECIFICATIONS (1) Description ( D PREFIX) ( A PREFIX) ( E SUFFIX) Control Voltage Range 3-32 VDC 90-280 Vrms 18-36 Vrms
Maximum Reverse Voltage -32 - -
Maximum Turn-On Voltage 3.0 VDC 90 Vrms 18 Vrms
Minimum Turn-Off Voltage 1.0 VDC 10 Vrms 4.0 Vrms
Typical Input Current 3.4-20 mA 2.0-4.0 mA 3 mA
Nominal Input Impedance [Ohms] 1500 Ohm 60 K Ohm 9.0 K Ohm
Maximum Turn-On Time [msec] (4) 1/2 Cycle 10 10
Maximum Turn-Off Time [msec] 1/2 Cycle 40 40
GENERAL SPECIFICATIONS Description Parameters Dielectric Strength, Input/Output/Base (50/60Hz) 4000 Vrms
Minimum Insulation Resistance (@ 500 V DC) 109 Ohm
Maximum Capacitance, Input/Output 8 pF
Ambient Operating Temperature Range -40 to 80 °C
Ambient Storage Temperature Range -40 to 125 °C
Weight (typical) 3 oz (86.5g)
Encapsulation Thermally conductive Epoxy
Terminals Screw and saddle Clamps Furnished, Unmounted
Recommended Terminal Screw Torque Range: 6-32 Screws - 10 in/lbs. 8-32 & 10-32 Screws - 20 in. lbs. (Screws dry without grease)
Fastons: Single pair (up to 25A) Double pair* (50A model only) *Caution: User must connect to both pairs
GENERAL NOTES 1) All parameters at 25°C unless otherwise specified.
2) Off-State dv/dt test method per EIA/NARM standard RS-443, paragraph 13.11.1
3) Heat sinking required, for derating curves see page 3.
4) Turn-on time for Random turn-on versions is 0.02 msec (DC Control Models)
AGENCY APPROVALS UL E116949
CSA LR81689
VDE 10143 UG (Not Applicable: -B and 4D) Rev. 030609
44
Published by the Health & Safety Executive 12/16
Spontaneous ignition of gas turbine lubricants at temperatures below their standard auto-ignition temperatures
RR1076
www.hse.gov.uk
There have been a number of incidents resulting in lubricating oil leaking in offshore gas turbine enclosures which could ignite if they came into contact with hot surfaces below their Auto Ignition Temperature (AIT). To assess the risk of auto-ignition, standard minimum AITs are used. However, AITs under industrial conditions are difficult to calculate and can be less than these standard values.
This report describes research using a Spontaneous Combustion Calorimeter developed to study spontaneous ignition. Preliminary tests were done for a range of process conditions that can influence minimum AITs for a number of gas turbine lubricating oils. These showed that ignition can occur at temperatures well below the standard minimum AIT. This indicates that if manufacturers rely on standard AITs at the design stage of gas turbines and enclosures, it may lead to a system that is likely to increase the ignition probability of any flammable release. To confirm these findings, further tests would be needed over a wider temperature range and under conditions which more closely represent the conditions in gas turbine enclosures.
Until AITs under industrial conditions are understood and addressed in design criteria, dutyholders will need to err on the side of caution in identifying and adequately controlling potential ignition sources.
This report and the work it describes were funded by the Health and Safety Executive (HSE). Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy.